Scheduling

To run DAGs automatically, you need to run the dagu scheduler process on your system. You can also use a cron expression generator for your scheduler calculation.

Cron Expression

You can specify the schedule with cron expression in the schedule field in the config file as follows:

schedule: "5 4 * * *" # Run at 04:05
steps:
  - name: scheduled job
    command: job.sh

Or you can set multiple schedules:

schedule:
  - "30 7 * * *" # Run at 7:30
  - "0 20 * * *" # Also run at 20:00
steps:
  - name: scheduled job
    command: job.sh

You can also specify a cron expression to run within a specific timezone. See list of tz database timezones:

schedule: "CRON_TZ=Asia/Tokyo 5 9 * * *" # Run at 09:05 in Tokyo
steps:
  - name: scheduled job
    command: job.sh

Stop Schedule

If you want to start and stop a long-running process on a fixed schedule, you can define start and stop times as follows. At the stop time, each step's process receives a stop signal:

schedule:
  start: "0 8 * * *" # starts at 8:00
  stop: "0 13 * * *" # stops at 13:00
steps:
  - name: scheduled job
    command: job.sh

You can also set multiple start/stop schedules. In the following example, the process will run from 0:00-5:00 and 12:00-17:00:

schedule:
  start:
    - "0 0 * * *"
    - "12 0 * * *"
  stop:
    - "5 0 * * *"
    - "17 0 * * *"
steps:
  - name: some long-process
    command: main.sh

Restart Schedule

If you want to restart a DAG process on a fixed schedule, the restart field is also available. At the restart time, the DAG execution will be stopped and restarted again:

schedule:
  start: "0 8 * * *"    # starts at 8:00
  restart: "0 12 * * *" # restarts at 12:00
  stop: "0 13 * * *"    # stops at 13:00
steps:
  - name: scheduled job
    command: job.sh

The wait time after the job is stopped before restart can be configured in the DAG definition as follows. The default value is 0 (zero):

restartWaitSec: 60 # Wait 60s after the process is stopped, then restart the DAG
steps:
  - name: step1
    command: python some_app.py

Queue Processing

The scheduler also manages the DAG execution queue. When a DAG is scheduled or manually enqueued, it's added to a queue system that controls concurrent execution based on configured limits.

Queue Behavior:

• DAGs are processed based on their queue assignment and concurrency limits
• Each DAG can be assigned to a named queue (defaults to the DAG name)
• Concurrent execution is controlled by maxActiveRuns settings
• The queue system can be disabled globally or per-DAG

Example with Queue Control:

name: batch-processor
schedule: "0 * * * *"   # Run every hour
queue: "batch"          # Use the "batch" queue
maxActiveRuns: 2        # Allow max 2 concurrent runs
steps:
  - name: process
    command: process_batch.sh

If the queue system is disabled (DAGU_QUEUE_ENABLED=false or in config), scheduled DAGs will run immediately without queue control.

Run Scheduler as a Daemon

The easiest way to make sure the process is always running on your system is to create the script below and execute it every minute using cron (you don't need root account in this way):

#!/bin/bash
process="dagu scheduler"
command="/usr/bin/dagu scheduler"

if ps ax | grep -v grep | grep "$process" > /dev/null
then
    exit
else
    $command &
fi

exit

Skip Successful Runs

To prevent redundant executions of scheduled DAGs, you can set skipIfSuccessful flag to true. When enabled, Dagu will check if the DAG has completed successfully since its last scheduled time. If it has, the current run will be skipped:

schedule: "0 */4 * * *"  # Run every 4 hours
skipIfSuccessful: true    # Skip if already succeeded since last schedule
steps:
  - name: resource-intensive-job
    command: process_data.sh

This is particularly useful for resource-intensive tasks where unnecessary re-runs should be avoided. Note that this only affects scheduled runs - manual triggers will always execute regardless of this setting.

For example, with the above configuration:

• If the DAG runs successfully at 04:00
• And someone triggers it at 05:00
• The run will be skipped because there's already a successful run since the last schedule
• The next run will occur at the next scheduled time (08:00)

The default value is false, meaning DAGs will run on every schedule by default.